One of the common roadblocks to successful model development or process implementation is how most leadership insist on immediate results and return on their investment in analytics. Instead of taking time to design and develop each analytic model, it is an all too familiar scene when management demands implementation in impractical time frames.

With predictive analytics, many organizations eschew in-depth assessment and custom tailored design, instead going for instant forecasts and summaries that result in insights that are “nice to know”, but do not have enough value to support decisions that make an impact or meet organizational goals.

Several considerations have to be addressed before implementation of analytics in an organization. Many practitioners fail to look into these issues before jumping headfirst into exploring the available data. This often leads to less favorable results with process implementation or development, but this is usually not because of limitations with analytics, instead being a failure in soft skills.

Below are some key issues that should be addressed by practitioners:

• Team Abilities
  • Are members fully supportive of the value of a solid design and plan before development? Do they agree with prioritizing project definition and strategic implementation in order to see the project to success?
• Leadership Buy-in
  • Is leadership motivated and fully on-board with analytics? Are they concerned with the organization’s analytic capabilities compared to industry competition or do they view the practice as having a more abstract function?
• Team Make-up
  • What are the motivations, goals, experience and concerns of the team that will be affected by or contribute to analytics? Are any members resistant to the shift? Will major errors be made in preparing data and interpreting results if the team doesn’t have oversight from professional modelers?
• Organizational Culture
  • Is the organization driven mostly by leadership choices and experience, or are they open to evidence-based decisioning, such as the results of A/B testing?
• Baselines and Targets
  • Are there established baselines for present organizational performance, as well as goal performance and how it will impact the business? If this data is not available, how will the organization define or measure whether the analytic project is a success?
• Recommendations
  • Is the organization willing and capable of following through on recommendations resulting from the predictive model?